Date: Mon, 19 Nov 2007 11:51:15 +0100 (CET)
Attached to this mail is a patch that introduces the --data-urlencode option,
as previously discussed. I do have some further thoughts about it that I'd
welcome feedback on! Here's the current usage:
The 'name' part is not encoded, only the content part (which can be loaded
from a file). Will we need any further tweaks, like allowing the entire given
string to be encoded in case someone wants to post something that isn't in
Currently, the name part just cannot contain a '=' or '@' letter as they will
be treated as separators between the name and the following data. Should we
allow a way to quote them or just live with that limitation?
This whole urlencode thing is introduced only to make it easier to write
scripts etc that do posts, as it removes the need for the scripts to do the
somewhat complicated procedure of url-encoding the data. Even though very few
people commented on this thread previosuly, I do believe this option will be
appreciated by the crowds once we provide it...
So, what are your comments and suggestions?
-- Commercial curl and libcurl Technical Support: http://haxx.se/curl.html
- TEXT/PLAIN attachment: data-urlencode.patch